This change attempts to only to override the default `pointer-events` behaviour for the `quo/keypair` component when the component uses the `selector` action. Now when the `quo/keypair` component is rendered with a selector, only the top-level view of the component will be pressable. And when the `quo/keypair` component is rendered without a selector it will allow for the children of the component to be pressable too.
Removes code integrating with the signal wakuv2.peerstats. The response from
this signal is rather large and it can sometimes happen in short bursts.
- Fixes https://github.com/status-im/status-mobile/issues/20176
We were using it to keep track of the number of peers in real-time. We now only
fetch the peer count in two instances (synchronously inside rn/use-mount):
- When the user opens the legacy screen Settings > Advanced > Peers stats.
- When the user opens the visibility status dropdown in Settings, because the
peers count will be used when the user chooses "Offline" (status inactive).
Note: we don't send anymore the peers count when the user shakes the phone. We
assume this value is already available in :peers-summary (which will be shared
by the user).
status-go still publishes the signal, we just ignore it to remove the
performance overhead of transforming to cljs and mutating the app-db (thus
interrupting the queue from more important stuff).
Steps to test:
- Verify user can see peers count in Settings > Advanced > Peers stats
- Verify changing the visibility status of user Alice in the Settings screen
leads to the desired effect to the other party seeing Alice's visibility
status.
This commit adds syncing saved addresses across paired devices whenever a new saved address is added or an existing one is edited or removed.
Signed-off-by: Mohamed Javid <19339952+smohamedjavid@users.noreply.github.com>
Fixes the scroll behavior in a community channel member list. The problem was
that the code wasn't enforcing the correct height.
Partially solves https://github.com/status-im/status-mobile/issues/20514
Steps to test: Long press on any community channel and press View channel
members and details. The user should be able to correctly scroll throughout the
entire list of members.
BigNumber instances are not comparable to other data types, except numbers and
strings (supported directly by the library). This commit makes = safe on
bignumbers (no exceptions should be thrown), even when comparing a bignumber
with a nil value.
There's another important benefit from having a custom comparator: performance,
because re-frame will be able to correctly compare data with bignumbers and thus
reuse cached subscriptions and also Reagent will be able to skip processing when
props are equal.
The IComparable interface was not changed. Similarly to core CLJS data types,
sorting heterogeneous collections throws (ignore nils).
```clojure
;; This throws with message: cannot compare 10 to foo
(sort ["foo" 10])
```
We could make sorting work with heterogeneous collections, but then we would
need to extend many core data types to play nice with BigNumber and that sounds
more like a hack, and it's also not a problem we have nowadays.
* Fix color of non-selected keypairs
* Fix pressable area in keypairs
* Make validation more robust
* Make confirm button always pressable
* Fix wrong sentence in errors
* Solve account creation by fixing a memoized callback
* Improve logging
* tweak: use support blur colors for action drawer icons
* tweak: support blur colors for drawer-top icon
* tweak: support blur colors for account list inside keypair component
* tweak: use blur colors for missing-keypair bottom sheet actions
* feat: remove feature-flag for key-pairs and accounts inside wallet settings
* fix: ensure key-pairs-and-accounts settings are visible
* tweak: display key-pair rename screen as full-screen modal
* tweak: animate key-pair scanning screen as modal
* fix: do not allow user to rename key-pair to an existing key-pair name
* fix: ensure we visualise the validation error for key-pair name that is too long
* fix: ensure we can see the recovery-phrase suggestions when testnet-mode is active
* tweak: show recovery-phrase errors when keyboard is hidden
* fix: ensure we do not re-hash password
* fix: ensure we call `on-close` function when successfully handling biometrics auth
* fix: ensure we clear error states when editing the recovery-phrase input
* fix: use blur styles for standard authentication password input
* chore: add labels for qr validation
* tweak: handle display import-qr error and allow for re-scanning
* tweak: update blur background for all bottom-sheets
* chore: add feature-flag for import-all key-pairs button for missing key-pairs
* tidy: remove unused variable
* fix: ensure layout for android is correct when entering seed-phrase in testnet-mode
* fix: ensure we return the updated db when updating the db with backup key-pair
* tweak: coerce nil to empty string
* tweak: change validation error message for short key-pair name
* tweak: hide the options icon for the default key-pair inside key-pairs and accounts settings
* chore(wallet): make recent tab fullscreen for empty view
* chore(wallet): hide advanced button on send page
* chore(wallet): hide estimated fees when no routes found
* chore(wallet): update emoji icon to edit
* chore(wallet): update key pair placeholder to use keypair example
* chore(shell): feature flag saved address option in scanner
After migrating from v1 to v2, some user names may be too long to fit in some
devices in the "Settings > Edit Profile > Name" setting. The solution was to
truncate in the UI the name to the max valid length in v2, which is 20
characters.
Other small improvements:
- Don't use entire Clojure maps as the :key in React. This is known to be slow
because it's expensive for Clojure to generally transform a data structure
into its string representation.
- Don't find the last item of a list using last and an equality between two
maps. This will force each iteration to call last, which in turn will iterate
over the entire list of items (we can think of the Clojure lazy seq as a
linked list). Instead, we now cache the index outside the loop and do a number
comparison.
- Don't pass entire props to style functions when not needed because
destructuring has a cost. I wouldn't personally worry about this 99% of the
time, but when there's only argument like in this PR where we were passing a
single blur? argument, it's preferable to not use a map.
Solves a problem encountered in unit tests, where instances of BigNumber (used
in the utils.money namespace) can't be compared with = or match?
If you try the following code you will always get false because they're
different instances in memory.
(= (money/bignumber 0) (money/bignumber 0))
We now extend BigNumber (from lib bignumber.js) so that their instances are
comparable (e.g. necessary to sort) and understand equality.
We also extend BigNumber to work with the match? directive. For example, this is
the output of comparing a nested map containing bignumbers (colored in the
terminal):
(match? {:a {:b {:c (money/bignumber 42)}}}
{:a {:b {:c (money/bignumber 43)}}})
FAIL in (equivalence-test) (at cljs$test$run_block (status-mobile/target/test/cljs-runtime/cljs.test.js:366:13)
expected: (match? {:a {:b {:c (money/bignumber 42)}}} {:a {:b {:c (money/bignumber 43)}}})
actual: {:a
{:b {:c (mismatch (expected "42") (actual "43"))}}}